Development of Safety Qualification Thresholds...
Transcript of Development of Safety Qualification Thresholds...
23 February 2007
Page 1 of 34
Development of Safety Qualification Thresholds and Their Use in Orally Inhaled and Nasal 1
Drug Product Evaluation 2
Douglas Ball, James Blanchard,† David Jacobson-Kram,
‡ Roger O. McClellan,
§ Timothy 3
McGovern,‡ Daniel L. Norwood,
¶ Mark Vogel,
# Ron Wolff,**
Lee Nagao
tt 4
5
Submitted to Toxicological Sciences for consideration as a Review article 6
7
8
To whom correspondence should be addressed. Safety Sciences, Groton Laboratories, Pfizer Global Research
and Development, Eastern Point Road MS 8274-1215; Groton, CT 06340; [email protected]
† Aradigm Corp., 3929 Pt. Eden Way, Hayward, CA 94545; [email protected]
‡ Office of New Drugs, Center for Drug Evaluation and Research, U.S. Food and Drug Administration, Silver
Spring, MD 20993; [email protected]; [email protected]
§ University of New Mexico Health Sciences Center and Toxicology and Risk Analysis, 13701 Quaking Aspen
Place NE, Albuquerque, NM 87111; [email protected]
¶ Boehringer Ingelheim Pharmaceuticals, Inc; [email protected]
# Safety Sciences, Pfizer Global Research and Development, 2800 Plymouth Road, Ann Arbor, MI 48105;
** Nektar Therapeutics; [email protected]
tt Drinker Biddle and Reath, LLP, 1301 K Street, NW, Suite 900, East Tower; Washington, DC, 20005; Phone:
202-230-5165. FAX: 202-230-5365; [email protected]
23 February 2007
Page 2 of 34
Abstract 9
Safety thresholds for chemical impurities and leachables in consumer products such as 10
foods and drugs have helped to ensure public health while establishing scientifically sound limits 11
for identification and risk assessment of these compounds. The Product Quality Research 12
Institute (PQRI) Leachables and Extractables Working Group, a collaboration of chemists and 13
toxicologists from FDA, industry, and academia, has developed safety thresholds for leachables 14
and extractables in orally inhaled and nasal drug products (OINDP), for application in United 15
States pharmaceutical submissions. The PQRI safety concern threshold (SCT) is 0.15 g/day, 16
and the qualification threshold (QT), 5 g/day. OINDP are important in the treatment of lung 17
diseases such as asthma and chronic bronchitis, as well as systemic diseases such as diabetes. 18
Analysis of extractables and minimization of leachables in OINDP are vital to ensuring the 19
quality and safety of the final product. It is expected that the thresholds developed by the PQRI 20
Leachables and Extractables Working Group will be used by both industry and regulators to 21
ensure and assess such quality and safety in OINDP applications. In this paper, we describe the 22
importance of the PQRI safety thresholds in the OINDP pharmaceutical development process; 23
the background and context of safety thresholds for consumer products; how these safety 24
thresholds were developed using well-established, robust databases and quantitative risk 25
assessment approaches; and how these thresholds can be applied in a pharmaceutical safety 26
qualification process, including FDA regulatory perspectives on the use of safety thresholds for 27
OINDP. 28
Keywords: qualification threshold, safety concern threshold, leachables, extractables, 29
inhalation, nasal, PQRI 30
31
23 February 2007
Page 3 of 34
I. Introduction 32
A. Why are extractables and leachables important in inhalation drug product 33
pharmaceutical development? 34
Extractables, as defined by the U.S. Food and Drug Administration (FDA), are 35
compounds that can be extracted from elastomeric components, plastic components, or coatings 36
of the container and closure system when in the presence of an appropriate solvent(s). 37
Leachables are compounds that leach from elastomeric or plastic components or coatings of the 38
container and closure system as a result of direct contact with the drug product formulation 39
(Food and Drug Administration, 1998 and July 2002). Extractables are therefore potential 40
leachables, and patients could be exposed to leachables. Leachables in orally inhaled and nasal 41
drug products (OINDP) are generally considered by pharmaceutical regulatory agencies to be of 42
significant safety concern (FDA, 1999) because for some OINDP, leachables could be delivered 43
directly to the diseased lung. For OINDP, toxicological and chemical assessments of leachables 44
and extractables are a critical part of the pharmaceutical development process. Further, 45
extractables and leachables can often impact on the marketing approval of a drug product. 46
OINDP include nasal sprays and inhalation sprays, metered dose inhaler (MDI) solutions and 47
suspensions, inhalation solutions, and dry powder inhalers (DPIs). These drug products are used 48
to mitigate the effects of indications such as asthma, emphysema, and allergy symptoms, and to 49
deliver medication directly to affected areas such as the nasal passages and lungs (local 50
delivery). Newer inhalation products, such as those for delivery of insulin to treat diabetes, seek 51
to deliver medication throughout the body, via the lungs (systemic delivery). 52
Each inhalation or nasal drug product, includes a container closure system that is a vital 53
part of the drug product as a whole, facilitating drug delivery and protecting the integrity of the 54
23 February 2007
Page 4 of 34
formulation. For example, the MDI consists of a solution or suspension formulation (drug 55
substance or active pharmaceutical ingredient; chlorofluorocarbon (CFC) or hydrofluoroalkane 56
(HFA) propellants to facilitate aerosol dose delivery; and surfactants, co-solvents and other 57
excipients to help stabilize the formulation), and a container closure system with various physical 58
components (metal canister to contain the pressurized formulation; a valve to meter the 59
formulation dose to the patient; elastomer components to seal the valve to the can and contain the 60
pressurized formulation; and an actuator/mouthpiece to facilitate patient self-dosing). The 61
formulation and container closure system are closely integrated in the MDI drug product, which 62
is generally true of all OINDP. The container closure systems for OINDP from which leachables 63
may appear can include not only primary packaging components, e.g., canisters for MDIs and 64
blister packages for DPIs, but also secondary packaging, such as overwraps or labels that are not 65
in direct contact with the drug formulation. 66
OINDP container closure systems components can consist of various elastomeric (rubber) 67
and polymeric (plastic) materials. These components can be seals, parts of MDI valves, 68
mouthpieces, etc. All such elastomeric and polymeric materials contain relatively low molecular 69
weight soluble organic chemical entities, either purposefully added to the materials during 70
synthesis, compounding, or fabrication (e.g., polymerization agents, fillers, antioxidants, 71
stabilizers, and processing aids), or present in the materials as a by-product of synthesis, 72
compounding, or fabrication (e.g., oligomers, additive contaminants, and reaction products). All 73
of these chemical entities have the capacity to leach into the formulation and be delivered to the 74
patient. Leachables and extractables represent a variety of chemical types and classes, and 75
leachables can be present in inhalation drug products at widely varying concentrations. 76
Extractables and leachables can also include certain structure types with known safety 77
23 February 2007
Page 5 of 34
implications, such as N-nitrosamines, polyaromatic hydrocarbons (PAHs), and 78
mercaptobenzothiozoles (FDA, November 1998 and July 2002). 79
B. Why do we need safety thresholds? 80
Modern analytical chemistry has enormous capability for analyzing extractables and 81
leachables. For example, Gas Chromatography/Mass Spectrometry (GC/MS) and Liquid 82
Chromatography/Mass Spectrometry (LC/MS), are capable of separating, identifying and 83
quantifying highly complex mixtures of organic chemical entities such as those produced from 84
Controlled Extraction Studies of OINDP container closure system components (see Figure 1) 85
(Norwood et al., 2005). These analyses are routinely accomplished at very high sensitivity, 86
easily detecting organic leachables at low µg/canister levels in MDIs (Norwood et al., 1995). 87
The available scientific literature would indicate that many hundreds, if not thousands, of 88
individual chemical entities that can appear as extractables/leachables could potentially be 89
detected, identified, and quantified at similar levels. 90
However, it is well established that there are levels at or below which organic chemical 91
entities in drug product represent no safety concern to patients. Therefore, the establishment of 92
safety thresholds that are protective of patients for OINDP leachables and extractables can be 93
justified and are believed to be necessary to limit unreasonable and extended evaluations of 94
chemicals present at levels that cannot harm patients. An efficient pharmaceutical development 95
process requires guidance on “how low to go” for the identification and quantification of 96
extractables and leachables. 97
23 February 2007
Page 6 of 34
98
99
Figure 1. A GC/MS extractables “profile” of an elastomer (Total Ion Chromatogram of a solvent extract). 100
5 . 0 0 1 0 . 0 0 1 5 . 0 0 2 0 . 0 0 2 5 . 0 0 3 0 . 0 0 3 5 . 0 00
5 0 0 0 0 0
1 0 0 0 0 0 0
1 5 0 0 0 0 0
2 0 0 0 0 0 0
2 5 0 0 0 0 0
3 0 0 0 0 0 0
3 5 0 0 0 0 0
4 0 0 0 0 0 0
4 5 0 0 0 0 0
5 0 0 0 0 0 0
5 5 0 0 0 0 0
6 0 0 0 0 0 0
6 5 0 0 0 0 0
7 0 0 0 0 0 0
7 5 0 0 0 0 0
8 0 0 0 0 0 0
8 5 0 0 0 0 0
9 0 0 0 0 0 0
9 5 0 0 0 0 0
1 e + 0 7
1 . 0 5 e + 0 7
1 . 1 e + 0 7
1 . 1 5 e + 0 7
1 . 2 e + 0 7
1 . 2 5 e + 0 7
1 . 3 e + 0 7
1 . 3 5 e + 0 7
T i m e - - >
A b u n d a n c e
T I C : 0 2 0 5 0 3 0 2 . D
23 February 2007
Page 7 of 34
-
To ensure quality and safety of OINDP through knowledge of amounts and types of 101
extractables and leachables, safety thresholds must be applied in conjunction with 102
pharmaceutical development “best practices,” which include safety assessment of potential 103
leachables throughout the development program. Such assessment should include at a minimum: 104
safety evaluation of supplier information during container closure component selection; safety 105
evaluation of potential leachables during Controlled Extraction Studies on OINDP components; 106
safety evaluation of leachables; and safety input to extractables specifications for routine quality 107
control of extractables. 108
C. The PQRI Leachables and Extractables Effort 109
In 2001, the Product Quality Research Institute (PQRI) initiated a project to develop 110
safety and analytical thresholds for leachables and extractables in OINDP.1 At the time there 111
was no regulatory guidance available for such thresholds. Furthermore, International Conference 112
on Harmonisation (ICH) thresholds for impurities are not applicable to leachables and 113
extractables in OINDP (ICH Q3A(R1), 2003; ICH Q3B(R2), 2003; ICH Q3C(R3), 2003). 114
The PQRI Leachables and Extractables Working Group, consisting of toxicologists and 115
chemists from industry, FDA, and academia, developed a safety concern threshold and a 116
qualification threshold for leachables; an analytical evaluation threshold for extractables and 117
leachables; processes for applying these thresholds; and best practices for selecting OINDP 118
container closure system components and conducting Controlled Extraction Studies, leachables 119
studies, and routine extractables testing. 120
1 Product Quality Research Institute: http://www.pqri.org
23 February 2007
Page 8 of 34
-
The PQRI safety concern threshold (SCT) is proposed to be 0.15 g/day, and the 121
qualification threshold (QT), 5 g/day. The SCT is the threshold below which a leachable would 122
have a dose so low as to present negligible safety concerns from carcinogenic and 123
noncarcinogenic toxic effects. The QT is the threshold below which a given non-carcinogenic 124
leachable is not considered for safety qualification (toxicological assessments) unless the 125
leachable presents structure-activity relationship (SAR) concerns. Below the SCT identification 126
of leachables generally would not be necessary. Below the QT leachables without structural 127
alerts for carcinogenicity or irritation would not require compound-specific risk assessment. 128
In 2006, the PQRI recommendations were submitted to the FDA for consideration in the 129
FDA’s development of regulatory recommendations for OINDP (PQRI, 2006). Safety 130
Thresholds and Best Practices for Extractables and Leachables in Orally Inhaled and Nasal Drug 131
Products. http://www.pqri.org). This paper describes, from a United States perspective, 132
historical background on the development and application of safety thresholds for consumer 133
products; the development and application of the PQRI safety thresholds for leachables in 134
OINDP; and FDA regulatory perspectives on the use of these safety thresholds. 135
I. Concept and History of Safety Thresholds 136
Over five hundred years ago Paracelsus made the astute observation that: “All substances 137
are poisons; there is none which is not a poison. The right dose differentiates a poison from a 138
remedy.” A corollary to this bit of wisdom is that for most or perhaps even all toxicological 139
effects, there exist thresholds: a dose below which an exposure imparts no risk. While most 140
toxicologists would likely agree with this principle, the means for calculating the threshold can 141
sometimes be controversial. In addition, for some adverse health endpoints, i.e., mutagenesis 142
and carcinogenesis, most regulatory agencies have assumed a lack of a threshold. Practically, 143
23 February 2007
Page 9 of 34
-
this means that any exposure results in some increased risk for mutation and/or cancer. 144
Nevertheless, it is often impossible to reduce human exposures to zero. 145
Two types of methodologies have evolved in risk assessment: one for calculating safe 146
exposure values for health effects thought to have thresholds, and a second for calculating 147
“virtually safe” exposure values for health effects thought to lack thresholds. In both cases, the 148
method involves extrapolation of agent-induced health effects in animals to human risk. For 149
drug development, methodologies for threshold-associated exposures are discussed in ICH Q3C, 150
Impurities: Guideline for Residual Solvents (ICH Q3C(R3), 2003). The guideline discusses a 151
method for calculating a PDE, “permissible daily exposure” to a residual solvent. Calculating 152
PDEs for mutagenic and carcinogenic substances is more complex. In this case, a “virtually safe 153
dose” can be established, using animal studies, as a dose that implies a negligible risk when 154
administered over a lifetime. A virtually safe dose has been defined somewhat differently by 155
different regulatory agencies. In general, however, it refers to lifetime exposures that increase 156
the risk of cancer by either one in one million or one in one hundred thousand. 157
These types of risk assessments are often used to calculate acceptable levels of 158
carcinogens in drinking water, air, and soil at hazardous waste sites. In order to perform such 159
risk assessments, data from rodent lifetime bioassays are required. Data from such studies are 160
often not available for impurities found in drug substances and drug products, and because of the 161
resource intensity and protracted nature of rodent lifetime bioassays, it is generally not practical 162
to test all potentially carcinogenic impurities in such an assay. This conundrum has led 163
toxicologists to devise the concept of the “threshold of toxicological concern” (TTC) (Kroes, et 164
al., 2004). These authors define the TTC as “a level of exposure for all chemicals, whether or 165
not there are chemical-specific toxicity data, below which there would be no appreciable risk to 166
23 February 2007
Page 10 of 34
-
human health.” The concept of a TTC appears to have originated with FDA’s threshold of 167
regulation for indirect food additives (FDA, 1995). 168
In general, an impurity exposure level of 1.5 µg/person/day is considered an acceptable 169
threshold below which further qualification for genotoxicity/carcinogenicity concerns would not 170
be required. The qualification threshold was originally developed as a “threshold of regulation” 171
by the Center for Food Safety and Applied Nutrition (CFSAN) at the FDA for food contact 172
substances and was further standardized by the CFSAN/FDA in a companion guidance document 173
for food contact substances (FDA, 1995 and April 2002). Substances with no known cause for 174
concern that may migrate into food are exempted from regulation as a food additive if present at 175
daily dietary concentrations at or below 0.5 parts per billion, corresponding to 1.5 µg/person/day 176
based on a total daily consumption of 3 kg of solid and liquid foods. The threshold is an estimate 177
of daily exposure expected to result in an upper bound lifetime risk of cancer of less than 10-6
, 178
considered a “virtually safe dose”. The initial CFSAN/FDA analysis was based on an 179
assessment of 343 carcinogens from a Carcinogenic Potency Database (CPDB) and was derived 180
from the probability distribution of carcinogenic potencies of those compounds (FDA, 1995; 181
Gold et al., 1984; Rulis, 1992). Subsequent analyses of an expanded database of more than 700 182
carcinogens further confirmed the threshold (Fiori and Meyerhoff, 2002). Additional analysis of 183
subsets of highly potent carcinogens suggested that a threshold of 0.15 µg/day, corresponding to 184
a 10-6
lifetime risk of cancer, may be more appropriate for chemicals with structural alerts for 185
potential genotoxicity (Kroes et al., 2004). Some structural groups including aflatoxin-like-, N-186
nitroso-, and azoxy-compounds were identified to be of extremely high potency and are excluded 187
from the threshold approach. 188
23 February 2007
Page 11 of 34
-
United States federal regulatory agencies such as the EPA and FDA typically use a 10-6
189
lifetime risk of cancer to determine “acceptable” risk from chemical exposures, although higher 190
risk levels are accepted under certain circumstances, namely for active pharmaceutical 191
ingredients from which a benefit may be derived. This level of exposure is expected to produce 192
a negligible increase in carcinogenic risk based on the analysis of the CPDB. Additionally, this 193
threshold is considered to be low enough to ensure that the presence of an unstudied compound 194
that is below the threshold will not significantly alter the risk/benefit ratio of a drug product, 195
even if the impurity is later shown to be a carcinogen. 196
In developing the SCT and QT, the PQRI Working Group carefully considered the 197
assumptions, approaches and, where possible, exposure data used in developing the threshold of 198
regulation for food additives, the TTC, and ICH thresholds for impurities. The Working Group 199
took a relatively conservative approach in the assumptions applied to development of the SCT 200
and QT. These assumptions were in many cases different than those used in development of 201
these other thresholds, taking into account the possibility of mixtures of leachables, and a real 202
potential for the presence of carcinogenic leachables. Further, unlike impurities, which are 203
associated with drug substance or drug product, leachables are not drug related and could possess 204
different toxic characteristics. The approaches to and assumptions made in establishing the SCT 205
and QT are explained below. 206
III. Safety Thresholds for Leachables in OINDP 207
Because leachables originate in the container closure system rather than the synthetic 208
pathway, the PQRI SCT and QT are based solely on µg/day intake of leachables, unlike ICH 209
thresholds for drug product impurities, which are linked to the dose of the active pharmaceutical 210
ingredient. The PQRI thresholds were defined in relation to estimated safe human inhalation 211
23 February 2007
Page 12 of 34
-
exposures for sets of chemicals assessed for different toxicity endpoints, an approach similar to 212
that used by others to determine thresholds for orally ingested substances (Blackburn et al., 213
2005; Cheeseman et al., 1999; Dolan et al., 2005; Fiori and Meyerhoff, 2002; Kroes et al., 2000; 214
Kroes et al., 2004; FDA, 1995; Munro et al., 1996; Rulis, 1992). 215
A. Derivation of the Safety Concern Threshold (SCT) 216
The SCT is based on carcinogenicity endpoints because carcinogenic effects occur at 217
lower intakes than those associated with noncarcinogenic toxicity. This was previously 218
demonstrated for orally ingested compounds, including those with potent neural, reproductive, or 219
endocrine toxicity (Kroes et al., 2000). Our analysis confirms that genotoxic carcinogenicity is a 220
concern at lower doses of inhaled compounds than acute respiratory irritation, or chronic 221
respiratory and systemic toxicity. 222
The Working Group based the SCT on the potencies of genotoxic carcinogens (i.e, 223
positive for mutations in Salmonella, SAL+) in the Carcinogenic Potency Database (CPDB), 224
which expresses carcinogenic potency as the TD50, the daily dose inducing a particular tumor 225
type in half of the animals that otherwise would not develop the tumor over a lifetime. Human 226
10-6
risk specific doses were estimated by linear extrapolation from the TD50 (Blackburn et al., 227
2005; Cheeseman et al., 1999; Fiori et al., 2002; Krewski et al., 1990; Kroes et al., 2004; Rulis, 228
1992). As expected (Cheesman et al., 1999), SAL+ carcinogens are more potent carcinogens 229
(Figure 2). The SAL+ compounds are also appropriate for establishing a threshold for structural 230
identification because structural alerts are more predictive for SAL+ than for nongenotoxic 231
carcinogens (Benigni and Zito, 2004). Furthermore, most known human carcinogens are 232
genotoxic (Bartsch and Malaveille, 1989), and the assumption of linear extrapolation of cancer 233
risk is more appropriate for SAL+ compounds than for nongenotoxic compounds, which may 234
23 February 2007
Page 13 of 34
-
exhibit mechanism-based thresholds for tumorigenesis. Too few inhalation studies were 235
represented in the CPDB to establish a threshold based solely on inhalation data. However, the 236
potency of carcinogens tested by inhalation mirrors that of those tested by all routes (Figure 2); 237
thus data from all routes should be representative of inhalation carcinogens. 238
A 10-6
risk specific dose was used as a negligible carcinogenicity risk in our analysis, 239
consistent with the threshold of regulation for indirect food additives (FDA, 1995). The 240
relatively conservative 10-6
risk level was considered an appropriate starting point for 241
identification and evaluation of leachable impurities because it is not uncommon for there to be a 242
mixture of multiple leachable impurities with potential genotoxicity issues in an OINDP, and 243
there are examples of potent carcinogens found as leachables in OINDP. 244
Our estimates of human 10-6
risk specific doses included allometric scaling factors, based 245
on default body weights (70-kg human, 350-g rat, 30-g mouse) to the 0.75 power, used by the 246
U.S. Environmental Protection Agency (EPA, 1992). In U.S. pharmaceutical labeling, dose 247
metrics from carcinogenicity assays are typically scaled to body surface area on a mg/m2 basis 248
(default body weight to the 2/3 power). These two specific dose-scaling approaches result in 249
relatively small (<2-fold) differences in human dose estimates (FDA, 2005), considering that 250
carcinogen potencies range over several orders of magnitude. Because applying multiple 251
conservative assumptions can unrealistically overestimate carcinogenic risk (Gaylor et al., 1993), 252
the Working Group used a central estimate of risk rather than an upper-bound 95% risk estimate, 253
and used the geometric mean of potencies from rats and mice rather than using the most sensitive 254
species. Based on these assumptions, the median human equivalent 10-6
risk specific dose for 255
SAL+ carcinogens in the CPDB is 0.36 µg/day, and the median excess cancer risk at the SCT of 256
0.15 µg/day is 0.41 × 10-6
. If <20% of random chemicals are genotoxic carcinogens (Fung et al., 257
23 February 2007
Page 14 of 34
-
1995; Sawatari et al., 2001), <7% of all compounds would exceed 10-6
increased cancer risk at 258
lifetime intakes <0.15 µg/day. Thus, a leachable below the SCT is unlikely to have a lifetime 259
excess cancer risk >10-6
, and identification of leachables below this threshold generally would 260
not be necessary. 261
10-6
Carcinogenicity Risk - CPDB Data
0%
20%
40%
60%
80%
100%
0.0001 0.001 0.01 0.1 1 10 100
10-6
Risk Specific Dose (µg/day)
Cu
mu
lati
ve
Pe
rce
nt
.
All SAL negative (N=178)
All SAL positive (N=276)
Inhalation SAL neg (N=12)
Inhalation SAL pos (N=27)
262
Figure 2. Carcinogenic potency of genotoxic (SAL-positive) and non-genotoxic (SAL-negative) 263
carcinogens from the Carcinogenic Potency Data Base (CPDB) from studies conducted in 264
rodents by inhalation or all routes combined. 265
266
B. Derivation of the Qualification Threshold 267
The Qualification Threshold is based primarily on regulatory agency "reference doses" 268
derived by applying safety factors to no-observed-adverse-effect levels (NOAELs) for 269
noncarcinogenic toxicity. The Working Group analyzed 150 inhaled compounds with "Chronic 270
Reference Doses"(RfDs) established by the US EPA, "Minimum Risk Levels" (MRLs) 271
23 February 2007
Page 15 of 34
-
established by the US Agency for Toxic Substances and Disease Registry, or "Reference 272
Exposure Levels" (RELs) established by the California EPA. The median reference values were: 273
120 µg/day (10th
percentile = 1.5 µg/day) for chemicals with reference values based on 274
respiratory toxicity, and 1940 µg/day (10th
percentile = 5.0 µg/day) for chemicals with systemic 275
toxicity endpoints. Based on the large safety margins (~100-fold) incorporated in reference 276
doses, leachables at intakes <5 µg/day should pose negligible health risks. Compounds with 277
respiratory toxicity and inhalation reference values less than 5 µg/day are dominated by metals 278
and metal salts, and by reactive compounds with readily identifiable structural alerts for irritant 279
potential, such as aldehydes and isocyanates. 280
In developing the QT, the Working Group focused on the total inhaled dose, which 281
assumes 100% deposition, rather than on the actual dose deposited in the lungs. The 5 µg/day 282
QT therefore represents a small percentage - between only 1% and 6% - of the quantity of 283
particulate that individuals are normally inhaling. Note that these percentages would be even 284
smaller if comparison were made to, for instance, air concentrations equal to the National 285
Ambient Air Quality Standard for the respirable fraction (PM10), which is considered protective 286
of public health including sensitive sub-populations. 287
The Working Group also considered acute respiratory irritation in relation to the 288
Qualification Threshold, since potential airway irritation and bronchoconstriction are concerns 289
for impurities in OINDP (Shaheen et al., 1994). A useful metric of sensory irritation is the 290
RD50, the concentration of an irritant that decreases respiratory frequency by 50% in mice 291
(Alarie et al., 1980). A good correlation was reported between occupational threshold limit 292
values and the value of 0.03 × RD50 (Schaper, 1993). To estimate safe doses for respiratory 293
irritants, the Working Group calculated the µg dose at the RD50 concentration inhaled for 294
23 February 2007
Page 16 of 34
-
10 minutes divided by 1000 for a large set of chemicals (Schaper, 1993). The additional 30-fold 295
safety margin for this metric, compared to the value of 0.03 x RD50, should be sufficient to 296
account for sensitive populations such as asthmatics as illustrated in Table 1. The distribution of 297
RD50/1000 doses was quite similar to the distribution of NOAELs for chronic respiratory 298
toxicity. As with the chronic respiratory toxicants, compounds with RD50/1000 values below 299
the 5 µg/day were predictably irritant compounds such as aldehydes, isocyanates, and nitriles. 300
Pediatric populations were also considered. Children would be adequately protected by 301
the proposed QT since it is based upon inhalation references values that are intended to protect 302
essentially all people, including sensitive subpopulations such as children. When establishing 303
RfD and RfC values, the EPA identifies the NOEAL, lowest-observed-adverse-effect-level 304
(LOAEL), or benchmark dose or concentration and then divides this value by a series of 305
uncertainty factors, One uncertainty factor relevant to children accounts for variability in toxic 306
response among people, including highly sensitive subjects, such as children and elderly. This 307
intraspecies uncertainty factor usually has a value of 10. This factor can be equally divided into a 308
toxicokinetic variability component with a default value of 3.16 [i.e., (10)1/2
], and a 309
toxicodynamic variability component also with a default value of 3.16, assuming these 310
components act independently. 311
The intraspecies uncertainty factor of 10 and the associated subfactors of 3.16 have been 312
justified for children based upon multiple studies that have compared the clinical response to 313
pharmaceutical agents in children versus adults as well as the toxic response to chemical agents 314
in younger versus older animals.(Burin, 1999; Dourson, 2002) For example, the National 315
Academy of Sciences Committee on Pesticides in the Diets of Infants and Children reviewed 316
several human and animal studies and concluded that the 10-fold intraspecies uncertainty factor 317
23 February 2007
Page 17 of 34
-
was sufficient to protect infants and children (Bruckner, 2000). Furthermore, comparison of the 318
toxicokinetic data of 60 xenobiotics and the toxicodynamic data of 49 xenobiotics in adults, 319
children, and other groups has shown the composite 10-fold factor covers the great majority of 320
the population (Renwick, 1998). Further work is needed to determine whether the default 321
uncertainty factors offer adequate protection for children, especially for inhaled exposure to 322
gases and particles, and for children with lung disease, such as asthma or cystic fibrosis 323
The preponderance of current data indicate that a leachable below the Qualification 324
Threshold is unlikely to cause acute or chronic noncarcinogenic toxicity, and in the absence of 325
structural alerts for carcinogenicity or irritation should not require compound-specific risk 326
assessment. 327
Table 1. Bronchoconstrictor Concentrations in Asthmatics Relative to Occupational 328
Short Term Exposure Limits (STELs) and RD50 Values 329
Compound STEL
(mg/m3)
RD50/1000
(µg/m3)
Bronchoconstriction in
Asthmatics Reference
Nitrogen Dioxide 9.4 655 None at 753 µg/m3 Tunnicliffe et al., 1994
Sulfur Dioxide 13 523 Range = 666 to
10500 µg/m3
(20- to 1.2-fold below
STEL)
Rubinstein et al., 1990
Sulfuric Acid 3.0 NA None at 46 µg/m3 (65-
fold below STEL)
Some at 130 µg/m3 (23-
fold below STEL)
Avol et al., 1990
Formaldehyde 2.45 39 None at 3700 µg/m3 for 3
hr (1.5 × STEL)
Sauder et al., 1987
Toluene
Diisocyanate
0.14 4.8 Most at >14 µg/m3 (10-
fold below STEL)
A few at 7 µg/m3 ( 20-
fold below STEL)
O’Brian et al., 1979
STEL = Short term exposure limit; RD50 = Concentration decreasing respiratory rate by 50% in mice; NA = Not available.
330
23 February 2007
Page 18 of 34
-
331
IV. Application of Safety Thresholds in the Pharmaceutical Development Process for 332
OINDP 333
In general, the use of safety thresholds is considered acceptable in cases where adequate 334
supporting data are available; the Division of Pulmonary and Allergy Products (DPAP) in FDA’s 335
Center for Drug Evaluation and Research/Office of New Drugs has incorporated the application 336
of safety thresholds in the qualification process conducted for leachables and extractables in 337
OINDP over the last decade. It is recognized that the appropriate application of safety thresholds 338
has advantages, including a reduction in the unnecessary expenditure of animals, time, effort, 339
and money. This reduced expenditure may allow greater resources to be applied to drug 340
development areas that present more significant safety concerns. 341
Currently, no formal regulatory guidance on the safety evaluation of leachables and 342
extractables is available. In the absence of formal guidance, the DPAP developed an internal 343
practice which includes the following general approach: identification of the compound, 344
determination of the maximum daily human exposure based upon the proposed product 345
specification (e.g., amount of specific leachable in drug product, typically in units of ppm or 346
μg/canister), conduct of a structure activity relationship (SAR) assessment for 347
genotoxic/carcinogenic potential through use of published lists (Ashby et al., 1989; Tennant and 348
Ashby, 1991) or software programs such as DEREK (Deductive Estimation of Risk from 349
Existing Knowledge, www.chem.leeds.ac.uk/luk/) or Multicase (www.multicase.com), and 350
review of available toxicology/safety data bases or conduct of toxicology studies as deemed 351
necessary, e.g., 14-90 day general toxicology, genetic toxicology. The final safety assessment 352
23 February 2007
Page 19 of 34
-
should be based on a consideration of the maximum expected daily human exposure, the 353
intended patient population, and the anticipated duration of use. 354
Specific considerations for the safety assessment of leachables and extractables in 355
OINDP include three primary components: systemic toxicity, local toxicity of the respiratory 356
tree, and mutagenic/carcinogenic potential. The DPAP identified a safety threshold for systemic 357
and local toxicity parameters based on an evaluation of the US EPA’s IRIS and HEAST 358
databases (Risk Assessment Information System, Chemical Specific Toxicity Values. Database 359
contains information taken from the United States Environmental Protection Agency’s (EPA’s) 360
Integrated Risk Information System (IRIS), the Health Effects Assessment Summary Tables 361
(HEAST). www.risk.lsd.ornl.gov/tox/tox_values.shtml) and concluded that there is no significant 362
safety concern for inhaled chemicals with a maximum expected daily dose of 5 µg/day (100 363
ng/kg) or less. Therefore, no further toxicity data are needed in most cases when the maximum 364
expected daily human exposure is below the stated threshold. 365
The safety threshold for systemic toxicity was derived from an evaluation of 36 366
chemicals listed in the EPA database that were administered via the inhalation route and were 367
associated with systemic toxicity. The presumed safe dose derived from the reference 368
concentrations (RfC’s) for these compounds were all 100 ng/kg with three exceptions; the 369
three exceptions had a “safe” dose of 80 ng/kg. Considering the large safety factors (1,000 – 370
10,000) that are incorporated into the calculation of RfCs, a safety threshold of 100 ng/kg was 371
considered reasonable. The safety threshold for respiratory toxicity was derived from an 372
evaluation of 20 chemicals with inhalation data that produced respiratory tract toxicity. All but 4 373
of these chemicals had presumed safe daily inhalation exposures greater than 100 ng/kg based on 374
the RfCs, even after incorporation of large safety factors (300-1,000). Of note, all of the 375
23 February 2007
Page 20 of 34
-
compounds with a presumed safe dose less than 100 ng/kg had a structural alert associated with 376
respiratory irritation. These structural alerts include isocyanates, aldehydes, organic acids, 377
strained heterocyclic rings, and halogenated aromatic rings. 378
As described earlier, subsequent data analyses conducted by the PQRI Working Group 379
expanded the database evaluation to include the ATSDR and CAL EPA databases. The Working 380
Group also concluded that a threshold of 5 µg/day presented a negligible safety concern for non-381
carcinogenic effects, and recommended a qualification threshold of 5 µg/day (100 ng/kg/day, 50 382
kg person). Thus, the threshold recommended by the Working Group is in agreement with 383
current DPAP practice. 384
There are some exceptions to the use of the above described threshold approach for 385
leachables and extractables in inhalation products, and these include compounds identified as 386
respiratory irritants and sensitizers, and those that present a known or suspected genotoxic or 387
carcinogenic potential. With regard to respiratory irritants and sensitizers, chemicals should be 388
evaluated for structural alerts associated with irritation or sensitization. This determination is 389
especially important when considering the indicated population for a given product. Most 390
inhalation products approved to date are indicated for treatment of pulmonary diseases such as 391
asthma. These patients are already considered to have a compromised respiratory function and 392
may be more sensitive to the effects of irritants or sensitizers. If a compound is considered to 393
have an irritant or sensitizing potential, patient risk should be assessed on a case-by-case basis 394
after evaluating the available information for the specific compound. Additionally, the clinical 395
experience with the drug product should be evaluated for evidence of any adverse effects. If no 396
concern is identified for irritancy or sensitization, the safety qualification threshold for systemic 397
and local toxicity of 5 µg/day is appropriate. For anticipated clinical exposures greater than 5 398
23 February 2007
Page 21 of 34
-
µg/day, safety qualification should be conducted for systemic and local toxicity, as described 399
later in this section. 400
The DPAP currently has no formal policy or practice with regard to a safety threshold for 401
leachables or extractables with an identified or suspected genotoxic or carcinogenic potential. 402
As discussed earlier, the PQRI Working Group proposed a SCT of 0.15 µg/day, a threshold 403
derived from calculated risk-specific doses of genotoxic (SAL-positive) carcinogens from the 404
CPDB and considered to be a dose below which a leachable would present negligible concern for 405
adverse carcinogenic and noncarcinogenic effects. The proposed SCT for negligible 406
carcinogenic effects expands on the DPAP’s previous use of thresholds that focused primarily on 407
general toxicological effects. The PQRI proposal is, however, similar to that described 408
previously by FDA’s Center for Food Safety and Nutrition for the safety assessment of food 409
contact materials (FDA, 2002),
and a similar proposal has been made to support safety thresholds 410
for genotoxic impurities (Müller et al., 2006). The proposed approach is supported by a large 411
database, and the applied cancer risk of 10-6
is considered appropriate due to the nature of the 412
chemicals that are commonly encountered as leachables and to the lack of any benefit derived 413
from their presence. Notably, high potency carcinogens (e.g., nitrosamines, PAHs) are excluded 414
from this threshold approach. While the PQRI Working Group recommendation has not yet 415
been formally accepted by the FDA, the proposal is considered in DPAP’s safety evaluation of 416
suspected or known carcinogenic leachables and extractables. 417
Leachables and extractables are considered adequately qualified for genotoxic or 418
carcinogenic potential if they are demonstrated to produce negative results in genotoxicity and/or 419
carcinogenicity assays or, in cases where these data are not available, if they lack structural alerts 420
for these endpoints. When no concern for genotoxic or carcinogenic potential is identified, a 421
23 February 2007
Page 22 of 34
-
qualification threshold of 5 µg/day is appropriate in the absence of supporting general toxicology 422
data and an identified potential for respiratory irritation or sensitization. If a leachable or 423
extractable is a known or suspected genotoxin or carcinogen, appropriate tests should be 424
conducted or a rationale provided to alleviate this concern. Alternatively, consideration should 425
be given towards reducing the drug product specification to a level associated with the PQRI 426
Working Group recommended SCT of 0.15 µg/day. If the compound is a known carcinogen, the 427
product specification should be set to a level associated with a cancer risk of less than or equal to 428
10-6
. 429
In general, when adequate safety data are available to support a proposed product 430
specification, conduct of a compound-specific risk assessment rather than a threshold-based 431
evaluation is recommended. It is anticipated that, in most cases, higher product specifications 432
would be supported from a safety standpoint when data are available to support a compound-433
specific risk assessment than would be through use of a threshold-based approach. In cases 434
where the initial safety evaluation demonstrates a lack of genotoxic/carcinogenic or airway 435
sensitization risk and the proposed specification is below the QT, additional risk assessment is 436
not necessary. 437
In cases where the maximum expected human exposure to a leachable is expected to 438
exceed DPAP’s qualification threshold of 5 µg/day, safety qualification for general toxicity 439
concerns may be provided through evaluation of published toxicity data or relevant regulatory 440
exposure limits such as US EPA air quality standards, or through the conduct of inhalation 441
toxicology studies of an appropriate duration (e.g., at least 90 days duration for chronic 442
indications). In some cases, data for chemicals with well-characterized toxicity profiles that 443
23 February 2007
Page 23 of 34
-
have a high degree of structural similarity to a leachable/extractable for which limited safety data 444
are available may be considered. 445
When toxicology data are used to support proposed product specifications, the most 446
relevant data should be considered. For example, if data are available from studies using both 447
oral and inhalation administration, the data derived from the inhalation studies are usually 448
considered to be most relevant. Safety margins for anticipated human exposures are calculated 449
based upon the NOAELs in animal studies. Generally, a 10-fold safety factor is applied for 450
cross-species extrapolation. In cases where safety data are only available from studies using the 451
oral route of administration, an additional 100-fold safety factor is applied, based on an 452
evaluation of a subset of the EPA HEAST database for which data were available following both 453
oral and inhalation administration. For this subset of chemicals, the presumed safe inhalation 454
doses derived from the RfCs were up to 100-fold lower than the identified reference doses 455
(RfDs) after conversion to a mg/kg/day dose. Therefore, a 1,000-fold safety factor (10 x 100) is 456
typically applied when using animal data from oral studies to support human inhalation use. 457
The PQRI Leachables and Extractables Working Group’s proposed process for safety 458
qualification of leachables incorporating use of the safety concern and qualification thresholds, is 459
shown in decision tree format in Figure 3. The decision tree supports FDA recommendations for 460
application of safety thresholds. Note that in some cases, decreasing the level of a leachable to 461
not more than the threshold can be simpler than providing safety data. Alternatively, adequate 462
data could be available in the scientific literature to qualify a leachable. If neither is possible, 463
additional safety testing should be considered. The studies considered appropriate to qualify a 464
leachable will depend on a number of factors, including the patient population, daily dose, and 465
duration of drug administration. 466
23 February 2007
Page 24 of 34
-
Figure 3. Decision Tree for Identification and Qualification. (a) If considered desirable, a 467
minimum screen (e.g., genotoxic potential) should be conducted. A study to detect point 468
mutations, in vitro, is considered an appropriate minimum screen. (b) If general toxicity studies 469
are desirable, one or more studies should be designed to allow comparison of unqualified to 470
qualified material. The study duration should be based on available relevant information and 471
performed in the species most likely to maximize the potential to detect the toxicity of a 472
leachable. On a case-by-case basis, single-dose studies can be appropriate, especially for single-473
dose drugs. In general, a minimum duration of 14 days and a maximum duration of 90 days 474
Is leachable
greater than
SCT?
Is leachable
unusually toxic, a
PNA, or a
nitrosamine?
Structure identified to extent that
SAR and literature assessment
can be performed?
Reduce to not
more than (less
than or equal to)
SCT?
Reduce to
not more than
QT?
Greater
than QT?
Any known human relevant
risks based on SAR
assessmentc and/or literature
search?
Reduce to
safe level?
Any clinically
relevant adverse
effects?
Lower thresholds may be
appropriate. The thresholds will
be dependant on the associated
risk. Establish acceptable level
with regulatory agency
No further
action
No further
action
No further
action
Reduce to safe
levelQualified
Establish alternate
acceptable level
with regulatory
agency
Consider patient population and duration
of use and consider conducting:
Literature-based risk assessments
Genotoxicity studies (e.g., point
mutation)a
General toxicity studies (one species,
usually 14 to 90 days)b
Other specific toxicity endpoints, as
appropriate
Risk assessment based
on SAR assessment,
literature search, and
other available
regulatory limits
Yes No
No Yes
Yes
Yes Yes
No
Yes
Yes
No
No
NoYes
NoNo
Yes No
Based on
assessment
Based on
assessment
Structure identified to extent that
SAR and literature assessment
can be performed?
23 February 2007
Page 25 of 34
-
would be considered appropriate. (c) For example, do known safety data for this leachable or its 475
structural class preclude human exposure at the concentration present? 476
477
478
The following case examples are presented to illustrate the safety qualification process. 479
Case 1. Bis-2-ethyl-hexyl sebacate (CAS RN 122-62-3) 480
The proposed product specification corresponded to a maximum daily human exposure of 481
9.1 µg/day (182 ng/kg/day for a 50 kg individual). The most relevant toxicity study identified 482
for this compound was a published chronic dietary study in rats in which a NOEL of 200 mg/kg 483
was observed. This dose corresponds to an acceptable human inhalation exposure of 0.2 484
mg/kg/day (200,000 ng/kg/day) after including a 1,000-fold safety factor for cross-species 485
extrapolation and for the use of data derived from oral administration to support inhalation use. 486
Even after incorporation of this safety factor, a greater than 1,000-fold safety margin for bis-2-487
ethyl-hexyl sebacate was present when comparing the acceptable human daily exposure to the 488
maximum anticipated human exposure associated with the proposed product specification. 489
Case 2. 4-toluenesulfonamide (CAS RN 70-55-3) 490
The proposed product specification corresponded to a maximum daily human exposure of 491
60 µg/day (1200 ng/kg/day). In this case, the sponsor provided no supporting rationale for the 492
proposed specification. However, a review of the literature indicated that only acute toxicity 493
data were available. In this case, the DPAP requested that the sponsor lower the product 494
specification to a level that corresponded to the Division’s safety qualification threshold of 5 495
µg/day or provide adequate toxicology data, such as a 90 day inhalation toxicity study, to support 496
their proposed product specification. 497
Case 3. Acenaphthene (CAS RN 83-32-9) 498
23 February 2007
Page 26 of 34
-
The proposed product specification corresponded to a maximum daily human exposure of 499
0.067 µg/day (1.33 ng/kg/day). As in the previous case, only acute toxicity data were available. 500
However, a State of Minnesota drinking water standard for acenaphthene is set at 400 µg/L (U.S. 501
EPA Office of Water, Federal-State Toxicology and Risk Analysis Committee (FSTRAC) 502
(November 1993). Summary of State and Federal Drinking Water Standards and Guidelines). 503
This standard corresponds to an acceptable daily inhalation exposure of 160 ng/kg assuming a 504
daily intake of 2 liters/day, a 50 kg BW, and incorporation of a 100-fold safety factor for the use 505
of oral data to support inhalation use. The calculated acceptable inhalation exposure for humans 506
provided a greater than 100-fold safety margin, when compared to the maximum daily human 507
exposure to acenaphthene through use of the drug product at the sponsor’s proposed 508
specification. In addition, the anticipated human exposure was well below the DPAP’s safety 509
qualification threshold of 5 µg/day. 510
Case 4. Nitrosamines 511
Nitrosamines can be present in certain rubber components of inhalation devices and are 512
known carcinogens. In one product, six species were identified at various levels. The 513
carcinogenic risk assessment was based on total nitrosamine exposure using the slope factor 514
calculated for NMDA. A maximum human daily exposure of up to 0.04 ng/kg, a level 515
associated with a cancer risk estimate of 10-5
, was accepted based on the overall risk-benefit 516
analysis and technological considerations, namely, the inability to manufacture rubber 517
components that do not potentially leach nitrosamine compounds. Although the Division 518
allowed a level that corresponded to a cancer risk estimate of 10-5
, DPAP encourages sponsors to 519
continue to reduce the potential for exposure and to strive to develop methods to eliminate the 520
presence of nitrosamine compounds. 521
23 February 2007
Page 27 of 34
-
As stated previously, issues arising from the need for adequate safety qualification for 522
leachables and extractables can often delay the approval of drug products. Therefore, some 523
consideration should be given to addressing these issues earlier in product development or 524
providing more substantive qualification information. Some ways in which the safety 525
qualification process can be improved include the selection of materials to limit the number and 526
level of potential leachables, the use of pre-extraction methods to lower potential exposures, and 527
the submission of a clear rationale to support the safety of proposed product specifications. 528
Often, the DPAP receives submissions containing only the proposed drug product specifications 529
for a given leachable or extractable with no supporting rationale. It must be stressed that 530
submission of an adequate supporting rationale often leads to a more efficient and expeditious 531
review of a submission, with regard to the safety qualification of leachables and extractables. 532
V. Conclusion 533
The safety thresholds for leachables in OINDP developed by the PQRI Leachables and 534
Extractables Working Group – a safety concern threshold of 0.15 μg/day and a qualification 535
threshold of 5 μg/day – provide scientifically sound guideposts for industry and regulators to 536
ensure, with a high degree of confidence, the safety of patients using these drug products. These 537
thresholds were developed taking into consideration accepted safety thresholds for indirect food 538
additives, impurities in drug substances, drug products, residual solvents, and genotoxic 539
impurities in drug products. A scientifically sound process using well-established databases, risk 540
assessment approaches, and literature information relevant to inhalation products was used to 541
derive the thresholds. The Working Group also developed a recommended safety qualification 542
process that incorporates application of the SCT and QT. This process supports FDA’s 543
recommended processes for application of safety thresholds. 544
23 February 2007
Page 28 of 34
-
The PQRI thresholds and qualification process were developed with FDA regulatory 545
support, and are expected to be used in pharmaceutical development programs and by regulators 546
assessing drug applications. These scientifically justified thresholds will have a significant 547
positive impact on the OINDP development and regulatory review process by removing 548
uncertainty in the development process and directly linking safety with quality of OINDP. 549
VI. Acknowledgements 550
The authors thank the Product Quality Research Institute and the International 551
Pharmaceutical Aerosol Consortium on Regulation and Science, and their member organizations, 552
for support of this work. 553
554
23 February 2007
Page 29 of 34
-
References 555
Alarie, Y., Kane, L., and Barrow, C. S. (1980). Sensory irritation: The use of an animal model to 556
establish acceptable exposure to airborne chemical irritants. In Toxicology: Principles and 557
Practices (Reeves, A. L., Ed.), pp. 48-92. John Wiley, New York. 558
Ashby, J., Tennant, R. W., Zeiger, E., and Stasiewicz, S. (1989). Classification according to 559
chemical structure, mutagenicity to Salmonella and level of carcinogenicity of a further 42 560
chemicals tested for carcinogenicity by the U.S. National Toxicology Program. Mutat. Res. 223, 561
73-103. 562
Avol, E. L., Linn, W. S., Shamoo, D. A., Anderson, K. R., Peng, R. C., and Hackney, J. D. 563
(1990). Respiratory responses of young asthmatic volunteers in controlled exposures to sulfuric 564
acid aerosol. Am. Rev. Respir. Dis. 142(2), 343-348. 565
Bartsch, H., and Malaveille, C. (1989). Prevalence of genotoxic chemicals among animal and 566
human carcinogens evaluated in the IARC Monograph Series. Cell Biol. Toxicol. 5(2), 115-127. 567
Benigni, R., and Zito, R. (2004). The second National Toxicology Program comparative exercise 568
on the prediction of rodent carcinogenicity: Definitive results. Mutat. Res. 566(1), 49-63. 569
Blackburn, K., Stickney, J., Carlson-Lynch, H. L., McGinnis, P. M., Chappell, L., and Felter, S. 570
P. (2005). Application of the threshold of toxicological concern approach to ingredients in 571
personal and household care products. Regul. Toxicol. Pharmacol. 43(3), 249-259. 572
Burin, G. J., Saunders, D. R. (1999). Addressing human variability in risk assessment – the 573
robustness of the intraspecies uncertainty factor. Regul. Toxicol. Pharmacol. 30, pp. 209-216. 574
23 February 2007
Page 30 of 34
-
Bruckner, J. V. (2000). Differences in sensitivity of children and adults to chemical toxicity: the 575
NAS panel report. Regul. Toxicol. Pharmacol. 31, pp. 280-285. 576
Cheeseman, M. A., Machuga, E. J., and Bailey, A. B. (1999). A tiered approach to threshold of 577
regulation. Food Chem. Toxicol. 37(4), 387-412. 578
Dolan, D. G., Naumann, B. D., Sargent, E. V., et al. (2005). Application of the threshold of 579
toxicological concern concept to pharmaceutical manufacturing operations. Regul. Toxicol. 580
Pharmacol. 43(1), 1-9. 581
Dourson, M., Charnley, G., Scheuplein, R. (2002). Differential sensitivity of children and adults 582
to chemical toxicity. II. Risk and Regulation. Regul. Toxicol. Pharmacol. 35, pp. 448-467. 583
Environmental Protection Agency (EPA) (1992). A cross species scaling factor for carcinogen 584
risk assessment based on equivalence of mg/kg0.75/day. Federal Register. 57, 24152-24173. 585
Fiori, J. M. and Meyerhoff, R. D. (2002). Extending the threshold of regulation concept: De 586
minimis limits for carcinogens and mutagens. Regul. Toxicol. Pharmacol. 35, 209-216. 587
Food and Drug Administration (FDA) (1995). 21 CFR Sec. 170.39 - Food additives: Threshold 588
of regulation for substances used in food-contact articles. Federal Register. 60(136), 36581-589
36596. 590
Food and Drug Administration (FDA), Center for Drug Evaluation and Research (CDER) 591
(November 1998). Draft Guidance for Industry: Metered Dose Inhaler (MDI) and Dry Powder 592
Inhaler Drug Products (DPI); Chemistry, Manufacturing and Controls Documentation. 593
23 February 2007
Page 31 of 34
-
Food and Drug Administration (FDA), Center for Drug Evaluation and Research (CDER), and 594
Center for Biologics Evaluation and Research (CBER) (1999). Guidance for Industry: Container 595
Closure Systems for Packaging Human Drugs and Biologics; Chemistry, Manufacturing, and 596
Controls Documentation. 597
Food and Drug Administration (FDA), Center for Food Safety and Applied Nutrition, Office of 598
Food Additive Safety (April 2002). Final Guidance for Industry: Preparation of Food Contact 599
Notifications for Food Contact Substances: Toxicology Recommendations. 600
Food and Drug Administration (FDA), Center for Drug Evaluation and Research (CDER) (July 601
2002). Guidance for Industry: Nasal Spray and Inhalation Solution, Suspension, and Spray Drug 602
Products; Chemistry, Manufacturing, and Controls Documentation. 603
Fung, V. A., Barrett, J. C., and Huff, J. (1995). The carcinogenesis bioassay in perspective: 604
Application in identifying human cancer hazards. Environ. Health Perspect. 103(7-8), 680-683. 605
Gaylor, D. W., Chen, J. J., and Sheehan, D. M. (1993). Uncertainty in cancer risk estimates. Risk 606
Anal. 13(2), 149-154. 607
Gold, L. S., Sawyer, C. B., Magaw, R., Backman, G. M., de Veciana, M., Levinson, R., Hooper, 608
N. K., Havender, W. R., Bernstein, L., Peto, R., Pike, M. C., and Ames, B. N. (1984). A 609
carcinogenicity potency database of the standardized results of animal bioassays. Environ. 610
Health Perspect. 58, 9-319. 611
Krewski, D., Szyszkowicz, M., and Rosenkranz, H. (1990). Quantitative factors in chemical 612
carcinogenesis: Variation in carcinogenic potency. Regul. Toxicol. Pharmacol. 12, 13-29. 613
23 February 2007
Page 32 of 34
-
Kroes, R., Galli, C., Munro, I., Schilter, B., Tran, L-A, Walker, R., and Würtzen, G. (2000). 614
Threshold of toxicological concern for chemical substances present in the diet: A practical tool 615
for assessing the need for toxicity testing. Food Chem. Toxicol. 38, 255-312. 616
Kroes, R., Renwick, A. G., Cheeseman, M., Kleiner, J., Mangelsdorf, I., Piersma, A., Schilter, 617
B., Schlatter, J., van Schothorst, F., Vos, J. G., and Würtzen, G. (2004). Structure-based 618
threshold of toxicological concern (TTC): Guidance for application to substances present at low 619
levels in the diet. Food Chem. Toxicol. 42, 65-83. 620
International Council on Harmonization (ICH) (2003). ICH Q3A(R1): Impurities in New Drug 621
Substances. Federal Register 68(68), 6924-6925. 622
International Council on Harmonization (ICH) (2003). ICH Q3B(R2): Impurities in New Drug 623
Products. Federal Register 68(220), 64628-64629. 624
International Council on Harmonization (ICH) (2003). ICH Q3C(R3): Impurities: Guideline for 625
Residual Solvents. Federal Register 68(219), 64352-64353. 626
Müller, L., Mauthe, R.J., Riley, C. M., Andino, M. M., De Antonis, D., Beels, C., DeGeorge, J., 627
De Knaep, A. G. M., Ellison, D., Fagerland, J. A., Frank, R., Fritschel, B., Galloway, S., Harpur, 628
E., Humfrey, C. D. N, Jacks, A. S., Jagota, N., Mackinnon, J., Mohan, G., Ness, D. K., 629
O’Donovan, M. R., Smith, M. D., Vudathala, G., and Yotti, L. (2006). A rationale for 630
determining, testing and controlling specific impurities in pharmaceuticals that possess potential 631
for genotoxicity. Reg. Tox. Pharm. 44, 198-211. 632
23 February 2007
Page 33 of 34
-
Munro, I. C., Ford, R. A., Kennepohl, E., Sprenger, J. G., Maier, A., and Dourson, M. (1996). 633
Correlation of structural class with no-observed-effect levels: A proposal for establishing a 634
threshold of concern. Food Chem. Toxicol. 34(9), 829-867. 635
Norwood, D. L., Prime, D., Downey, B. P., Creasey, J., Satinder, S. K., and Haywood, P. (1995). 636
Analysis of polycyclic aromatic hydrocarbons in metered dose inhaler drug formulations by 637
isotope dilution gas chromatography/mass spectrometry. J. Pharm. Biomed. Anal. 13(3), 293-638
304. 639
Norwood, D. L., Nagao, L., Lyapustina, S., and Munos, M. (2005). Application of Modern 640
Analytical Technologies to the Identification of Extractables and Leachables. Am. Pharm. Rev. 641
8(1), 78-87. 642
O'Brien, I. M., Newman-Taylor, A. J., Burge. P. S., Harries, M. G., Fawcett, I. W., and Pepys, J. 643
(1979). Toluene di-isocyanate-induced asthma. II. Inhalation challenge tests and bronchial 644
reactivity studies. Clin. Allergy. 9(1), 7-15. 645
Renwick, A. G., Lazarus, N. R. (1998). Human variability and noncancer risk assessment – an 646
analysis of the default uncertainty factor. Regul. Toxicol. Pharmacol. 27, pp. 3-20. 647
Rubinstein, I., Bigby, B. G., Reiss, T. F., and Boushey, H. A. Jr. (1990). Short-term exposure to 648
0.3 ppm nitrogen dioxide does not potentiate airway responsiveness to sulfur dioxide in 649
asthmatic subjects. Am. Rev. Respir. Dis. 141(2), 381-385. 650
23 February 2007
Page 34 of 34
-
Rulis, A. (1992). Threshold of regulation: Options for handling minimal risk situations. In Food 651
Safety Assessment (J. W. Finley, S. F. Robinson, and D. J. Armstrong, Eds.), pp. 132-139. 652
American Chemical Society Symposium Series 484, Washington, D.C. 653
Sauder, L. R., Green, D. J., Chatham, M. D., and Kulle, T. J. (1987). Acute pulmonary response 654
of asthmatics to 3.0 ppm formaldehyde. Toxicol. Ind. Health. 3(4), 569-578. 655
Sawatari, K., Nakanishi, Y., and Matsushima, T. (2001). Relationships between chemical 656
structures and mutagenicity: A preliminary survey for a database of mutagenicity test results of 657
new workplace chemicals. Industrial Health. 39, 341–345. 658
Schaper, M. (1993). Development of a database for sensory irritants and its use in establishing 659
occupational exposure limits. Am. Ind. Hyg. Assoc. J. 54, 488-544. 660
Shaheen, M. Z., Ayres, J. G., and Benincasa, C. (1994). Incidence of acute decreases in peak 661
expiratory flow following the use of metered-dose inhalers in asthmatic patients. Eur. Respir. J. 662
7(12), 2160-2164. 663
Tennant, R. W., and Ashby, J. (1991). Definitive relationships among chemical structure, 664
carcinogenicity and mutagenicity for 301 chemicals tested by the U.S. NTP. Mutat. Res. 257, 665
209-227. 666
Tunnicliffe, W. S, Burge, P. S, and Ayres, J. G. (1994). Effect of domestic concentrations of 667
nitrogen dioxide on airway responses to inhaled allergen in asthmatic patients. Lancet 344(8939-668
8940), 1733-1736. 669